diff options
| author | Zack Weinberg <zackw@panix.com> | 2018-07-12 12:25:12 -0400 |
|---|---|---|
| committer | Zack Weinberg <zackw@panix.com> | 2018-07-12 12:25:12 -0400 |
| commit | b3ee64bf08082367ea2116ea1fb2007e1a2896cd (patch) | |
| tree | efc30a65452525c93b2bee7c36343c5bdbc5abb6 /Makefile.am | |
| parent | 9f728b79718a8e8b371f48b75592beed638f9c2b (diff) | |
| download | libxcrypt-b3ee64bf08082367ea2116ea1fb2007e1a2896cd.tar.gz libxcrypt-b3ee64bf08082367ea2116ea1fb2007e1a2896cd.tar.bz2 libxcrypt-b3ee64bf08082367ea2116ea1fb2007e1a2896cd.zip | |
Add tests for get_random_bytes.
It turns out not to be *that* hard to exercise the fallback logic in
get_random_bytes, thanks to GNU ld's --wrap feature. There is also
some basic black-box testing of the get_random_bytes interface.
The change to randombytes.c itself ensures 100% predictable behavior
if get_random_bytes should ever be called with buflen zero.
Diffstat (limited to 'Makefile.am')
| -rw-r--r-- | Makefile.am | 10 |
1 files changed, 9 insertions, 1 deletions
diff --git a/Makefile.am b/Makefile.am index 51cba76..f99f842 100644 --- a/Makefile.am +++ b/Makefile.am @@ -127,7 +127,8 @@ check_PROGRAMS = \ test-crypt-pbkdf1-sha1 test-crypt-sha256 test-crypt-sha512 \ test-crypt-sunmd5 \ test-byteorder test-badsalt test-badsetting test-gensalt \ - test-short-outbuf + test-short-outbuf \ + test-getrandom-interface test-getrandom-fallbacks if ENABLE_OBSOLETE_API libcrypt_la_SOURCES += crypt-des-obsolete.c @@ -183,6 +184,13 @@ test_alg_sha1_LDADD = alg-sha1.lo test_alg_sha256_LDADD = alg-sha256.lo test_alg_sha512_LDADD = alg-sha512.lo +test_getrandom_interface_LDADD = randombytes.lo +test_getrandom_fallbacks_LDADD = randombytes.lo +if HAVE_LD_WRAP +test_getrandom_fallbacks_LDFLAGS = \ + -Wl,--wrap,getentropy -Wl,--wrap,getrandom -Wl,--wrap,syscall \ + -Wl,--wrap,open -Wl,--wrap,read -Wl,--wrap,close +endif # Every object file depends on crypt-symbol-vers.h and crypt-hashes.h, # which are generated files, so automatic dependency generation is not |